Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Erdős #1049: the rational-base contour of Zudilin's `(14,12,14;27)` forms

Definition
ErdosProblems_Erdos1049_RationalBaseContour

by willcook · Sep 27, 2026 · Mathlib c5ea003 (Lean v4.30.0)

erdos-1049formalizationlambert-series

The rational-base contour compares denominator and remainder rates of Zudilin's (14,12,14;27) forms after homogeneous evaluation. The submitted module contains the source declarations trigammaSeries, summable_trigammaTerm, trigammaSeries_nonneg, trigammaSeries_antitone, sum_le_trigammaSeries_sub, among others. Source topic: Erdős #1049: the rational-base contour of Zudilin's (14,12,14;27) forms.

Definition code
import Definitions.Def_ErdosProblems_Erdos1049_ZudilinHeightRegion
import Mathlib.Algebra.BigOperators.Group.Finset.Basic
import Mathlib.Analysis.PSeries
import Mathlib.Analysis.Real.Pi.Bounds
import Mathlib.Analysis.SpecialFunctions.Log.Basic
import Mathlib.NumberTheory.ZetaValues
import Mathlib.Tactic.FieldSimp
import Mathlib.Tactic.NormNum
import Mathlib.Tactic.Ring
import Mathlib.Topology.Algebra.InfiniteSum.Real

/-!
# Erdős #1049: the rational-base contour of Zudilin's `(14,12,14;27)` forms

The proof note `RationalBaseIrrationality_31_4_Proof.md` shows that the linear
forms of W. Zudilin, *Heine's basic transform and a permutation group for
q-harmonic series*, Acta Arith. 111 (2004) 153–164, in the direction
`(α₀,α₁,α₂;β) = (14,12,14;27)`, after homogenisation at a reduced rational base
`a/b`, prove that `F(a/b) = ∑_{m≥1} 1/((a/b)^m - 1)` is irrational whenever

`log b / log a < θ* := C₀ / C₁`,   `C₁ = 1091/2`,   `C₀ = 266 - (3/π²)(225 - J)`,

where `J = ∑_{i=1}^{13} (ψ₁(uᵢ) - ψ₁(vᵢ))` is the trigamma integral of Zudilin's
step function `ω` over its thirteen printed intervals `[uᵢ, vᵢ)` and
`ψ₁(x) = ∑_{k≥0} 1/(k+x)²`.

This module formalises the finite part of that theorem: the constants are
defined from the trigamma series, and Lean proves

* `eightyOne_twoHundredths_lt_zudilinContour : 81/200 < θ*` from the single
  `k = 0` term of each trigamma difference and `π > 3.14`;
* `zudilinContour_lt_half : θ* < 1/2` from `J < ψ₁(1/14) < 198`;
* membership of `31/4` and every power `(31/4)^r` in the contour region, and
  the inclusion of the older `81/200` region in it;
* exclusion of `3/2` from the contour region;
* the exact integer bookkeeping of the forms: `2M(a;b) = 2(266n² + 34n + 1)`,
  the degree step `d_{k+1} - d_k = 40n + 1 - k`, and the top degree
  `2 d_{b-1} = 1091n² + 81n + 2`.

No analytic step is formalised.  Zudilin's Lemma 7 (integrality in `ℤ[p]`),
his Lemma 2 (trigamma equidistribution of `φ(l)`), and the Archimedean
estimate stay in the ordinary proof note.  Membership in the contour region is
the exact hypothesis that note consumes; it is not by itself an irrationality
statement.
-/

namespace ErdosProblems.Erdos1049

/-! ## The trigamma series `ψ₁(x) = ∑_{k≥0} 1/(k+x)²` -/

/-- The series representation of the trigamma function.  Only this series is
used; the identification with `d²/dx² log Γ` is classical and not needed. -/
noncomputable def trigammaSeries (x : ℝ) : ℝ := ∑' k : ℕ, 1 / ((k : ℝ) + x) ^ 2













/-! ## The thirteen intervals and the constant `J` -/

/-- One interval's contribution `ψ₁(u) - ψ₁(v)`. -/
noncomputable def zudilinJTerm (u v : ℝ) : ℝ := trigammaSeries u - trigammaSeries v

/-- `J = ∫₀¹ ω(x) d(-ψ'(x))` over the thirteen intervals on which `ω = 1`
(Zudilin 2004, end of Section 5), written as the trigamma series. -/
noncomputable def zudilinJ : ℝ :=
  zudilinJTerm (1 / 14) (1 / 12) + zudilinJTerm (1 / 7) (1 / 6) +
    zudilinJTerm (3 / 14) (1 / 4) + zudilinJTerm (2 / 7) (1 / 3) +
    zudilinJTerm (5 / 14) (2 / 5) + zudilinJTerm (3 / 7) (7 / 15) +
    zudilinJTerm (1 / 2) (8 / 15) + zudilinJTerm (4 / 7) (3 / 5) +
    zudilinJTerm (9 / 14) (2 / 3) + zudilinJTerm (5 / 7) (11 / 15) +
    zudilinJTerm (11 / 14) (4 / 5) + zudilinJTerm (6 / 7) (13 / 15) +
    zudilinJTerm (13 / 14) (14 / 15)

/-- `C₁ = (α₀+α₁+α₂)β - (α₁²+α₂²+β²)/2 = 1091/2`, Zudilin's (25). -/
noncomputable def zudilinC1 : ℝ := 1091 / 2

/-- `C₀ = α₁²/2 + α₀α₁ + (β-α₂)(α₂-α₁) - (3/π²)(m² - J)` with `m = 15`,
Zudilin's (26). -/
noncomputable def zudilinC0 : ℝ := 266 - 3 / Real.pi ^ 2 * (225 - zudilinJ)

/-- The rational-base threshold `θ* = C₀/C₁ = 1/μ`, where `μ = C₁/C₀` is the
irrationality-exponent bound of Zudilin's Theorem 1. -/
noncomputable def zudilinContour : ℝ := zudilinC0 / zudilinC1

/-- The parameter region of the authored rational-base theorem: reduced bases
`a/b` with `log b / log a < θ*`.  Membership is the hypothesis the ordinary
proof consumes; it is not an irrationality statement. -/
def ZudilinContourRegion (a b : ℕ) : Prop :=
  Real.log b / Real.log a < zudilinContour

/-! ## Lower bound: `J > 77.6`, hence `θ* > 81/200` -/











/-! ## Upper bound: `J < ψ₁(1/14) < 198`, hence `θ* < 1/2` -/











/-! ## Membership and exclusion -/













/-! ## Exact integer bookkeeping of the forms at index `n` -/
















end ErdosProblems.Erdos1049
Source
Pinned Lean source: https://github.com/wcook04/plectis-erdos-lean/blob/c93c2e4dd86a2e317e0cb650ea244fee1afd59c2/ErdosProblems/Erdos1049/RationalBaseContour.lean#L46-L355 Paper exposition and prior-art bibliography: https://github.com/wcook04/plectis-erdos/blob/551bae6dc6e732cf85172d66323c8d2bc77ba962/paper/1049/erdos-1049-rational-base-lambert.tex#L1252-L1257 Correspondence: Erdős #1049 formal source; Zudilin is credited for the relevant rational-base Lambert-series method, without a novelty or all-rational-base claim.

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me